Dynamic Multi-Fault Diagnosis Based Root Cause Tracing for Liquid Storage Tanks Assembly Production Lines

• DOI: 10.20944/preprints202502.2259.v1
• Publication Date: 2025-03-03T03:06:51Z
• Authors: You Teng; Donghui Li; Hongkai Xue; Yunkai Zhou; Kefu Wang; Qi Wu
• Keywords: Tracing

Tracing the root causes of defective products in liquid storage tank (LST) production poses a formidable challenge due to complexity and coupling processes within production. In this paper, problem tracing cause LST which is caused by process parameter deviations or human operation errors during studied. A method that based on dynamic multi-fault diagnosis (DMFD) framework proposed. First, factorial hidden Markov model (FHMM) established depict state-transition product, where its status changes over time across processes. This achieved considering product state at each as outcomes inspection an observation state. Then, Viterbi algorithm employed solve transition matrix diagnostic FHMM. Finally, experimental verification carried out real assembly line, influence imperfect testing accuracy also considered. Experimental results show proposed achieves 100% rate for three typical quality issues.